Statement of Purpose

To discover solutions which achieve the goals of redundant e-mail, and share them with the world

The Goals of RedundantMail

There are two basic goals of RedundantMail:

• All messages need to be stored in multiple places, so that one server (or even point of presence) can go down, and users will still be able to receive and access their e-mail via the other, but will not receive duplicate copies of any one message
• Spread the load over multiple servers so that no one server is too heavily loaded (ideally this would include preferring servers near the end user)

Requirements

To find out what the requirements are for RedundantMail, see the RedundantMail Requirements [5].

Past Projects

In the past, there have been other projects to do similar things, but none of them seem to have been very widely used or successful. But some are listed here for reference anyway:

• Hardware solution [6]: This guy has solved the problem with two SCSI disks in a RAID array being accessed by two servers. The reason this is not good enough is because a) If a fire comes through, everything is dead (we want the possibility of multi-site redundancy), and b) the RAID device itself is still a single point of failure.
• MailWulf: Once upon a time, someone was trying to make one huge package which contained all the pieces for a redundant mail system. That project was called MailWulf [7]. Many useful ideas can still be found within its pages, especially in the White Paper [8].
• Design and Implementation of Highly Scalable E-mail Systems [9], the slides for a talk given by some important and knowledgeable guys about the relevant problems

Join our Forum [10].